Using SAS ® to Examine Aging Expectation (ERA-38) for Older Adults

Abbas S. Tavakoli · Julie Freelove-Charton

0 citations0/0 files stored

No readable copy stored yet — this record is metadata only.

No open-access location is known for this record yet.

Literature note Cite in a project Citation map

Aging and Gerontology Research Psychometric Methodologies and Testing Aging, Health, and Disability

Abstract

Our population is getting older every day. The numbers of older adults (ages 65 years and older) will increase to 72 million by 2030 [CDC &MCF 2007]. Participation in healthy aging behaviors, such as physical activity and a healthy diet, is one of the few known ways to prevent the onset of chronic disease and stem the rising costs of health care. Objectives: To test a multidimensional aging-expectation scale for older adults. Method This exploratory correlation study used a cross-sectional survey design. Participants were 459 older adults (65 or older) from the Greater Columbia metropolitan area of South Carolina. The 38-item Expectations Regarding Aging (ERA-38) survey was used to measure aging-expectations (Sarkisian, Hays, & Mangione, 2002). Instrument testing consisted of a series of factor analysis procedures including maximum likelihood, using squared multiple correlation and Promax rotation. Reliability was assessed using coefficient alpha estimates. Multiple imputations (MI) were used to replace the missing data for item. Results The missing data for the 38 items ranges from 1-30. MI method used to replace the missing value for each item. Thirty-eight items loaded on three factors of ageing expectation. The weighted variance explained by each factor was: 1) aging process (35%), 2) being isolated (33%), and 3) physical function (31%), for a total of 99% explained variance for all three factors. Factor loadings ranged from 0.32 to 0.75. Coefficient alpha estimates were 0.87 - 0.92 across the three subscales and 0.95 for the total scale. Examination of Pearson's correlation indicated each subscale was positively correlated with every other subscale and the total scale. The range of the Pearson’s correlation coefficients was from 0.60 to 0.89. Conclusion Factor analysis showed three factor for aging- expectation (ERA-38) and excellent reliability for total scales and subscales.